Role of imidazolium cations on the interfacial structure of room‐temperature ionic liquids in contact with Pt(111) electrodes

Abstract

Room‐temperature ionic liquids (ILs) have gained considerable attention as an important addition to conventional electrolytes because they exhibit large electrochemical windows and can reduce existing overpotentials in electrocatalysis. For the interfacial electrochemistry of ILs, a comprehensive understanding of molecular ions and the resulting electric double‐layer structures as a function of electrode potential is mandatory, but the structures are largely different from conventional electrolytes. For that reason, we have studied the interfaces of Pt(111) in contact with ILs using 1‐butyl‐3‐methylimidazolium [BMIM] and 1‐butyl‐2,3‐dimethylimidazolium [BMMIM] cations as well as bis(trifluoromethylsulfonyl)imide [NTf₂] anions. We applied vibrational sum‐frequency generation (SFG), where we interrogate vibrational bands from interfacial cations, anions, as well as interfacial water in situ and under potential control. Structuring of [NTf₂] anions and H₂O with electrode potential show hysteresis while a strong Stark tuning was absent. This indicates that the IL ions are oriented in the vicinity of the interface, without being directly adsorbed to the Pt(111) surface. Using the C‐H stretching band from CH groups at the imidazolium ring, the ring reorientation with electrode potential was qualitatively determined. The imidazolium ring reorients as a function of potential from a more parallel orientation to an upright orientation with respect to the interfacial plane. This leads to the formation of voids in the layered structure of ions at the interface, which can be then filled with H₂O as evidenced by an increased SFG intensity from O‐H stretching modes that are attributable to hydrogen‐bonded interfacial water. Comparing the responses of the ILs, particularly of [BMMIM][NTf₂], shows a compact structure and a significantly pronounced rearrangement of the imidazolium ring that can also facilitates better incorporation of H₂O and significantly affects the reorientation of [NTf₂] anions and, thus, causes a pronounced hysteresis with electrode potential.

INTRODUCTION

Room-temperature ionic liquids (ILs) have recently gained considerable attention in electrochemistry^[] and related fields where they are applied as electrolytes in batteries,^[] supercapacitors,^[] for the deposition of functional materials,^[] in electrocatalysis^[] or simply as lubricants with unique properties^[] just to mention a few of many possible applications.

ILs offer properties like excellent conductivity, a fairly low viscosity, and most notably, a wide electrochemical window^[] that outperforms many other electrolytes and makes ILs highly interesting for but not necessarily limited to electrochemical applications. To understand the unique properties of ILs in interfacial electrochemistry, molecular-level information from electrode/electrolyte interfaces is mandatory because the local structure of an IL at the interface dramatically affects the performance in a given application.^[]

Today, different models exist to describe electrode/electrolyte interfaces in classical electrolytes like water with inorganic salts such as NaCl. In particular, the Gouy-Chapman-Stern model is often applied and assumes a diffuse electric double layer (EDL) where ions follow a Boltzmann distribution and efficiently screen the electrode's surface charge. The Gouy-Chapman-Stern model predicts a linearly profile change in the compact layer and a monotonic exponentially decaying distance dependence of the electric potential in the diffuse layer.^[] Although the model is a good description of the EDL for many conventional electrolytes, it fails to describe the behavior of ILs at the electrode/electrolyte interface. That is because of following several reasons: ILs consist mostly of large organic cations with delocalized charges, in contrast to relatively compact ions such as alkali or halide ions. In addition, molecular ions of ILs coordinate weakly through Coulomb and dispersive interactions as well as by hydrogen bonding^[] and π-π-interactions,^[] which makes the application of known double-layer models difficult.^[] Recent studies of electrode surfaces immersed in IL revealed an EDL that exhibits a layered structure composed of alternating co-ion-rich and counterion-rich layers, where the first layer of ions at the interface is not able to fully compensate the surface charge, so that the next layer of interfacial ions gets accumulated with coions too. To overcompensate the net charge of these two layers and the surface charge density, a third and even a fourth layer can be present, but with a higher intermixing of cations and anions. In the layers at a larger distance from the electrode surface, the layering of anions and cations in an oscillating way gets less pronounced to a point where bulk-like properties are established. This distance dependence of the layering of IL ions at electrode interfaces at high electrode potentials is known as the lattice saturation effect.^[] Here, molecular dynamic simulations reveal a structural transition to a surface-frozen monolayer of closely packed counter ions in a Moiré-like structure and a more disordered layer at a larger distance from the electrode surface.^[] The presence of a Moiré-like structure was indeed supported by in situ STM experiments at the Au(111)/[BMIM][PF₆] interface.^[] Many studies exist so far that focus on the EDL structures of ILs on Au surfaces, but there are only a few works on Pt surfaces.^[] In recent works with ILs, in particular, the H₂O concentration has been shown to play a major role in reducing CO overpotentials.^[] [BMIM] and [BMMIM] showed a different reaction mechanism via imidazolium-2-carboxylate intermediate and electrostatic stabilization.^[] Under this aspect, a sufficient understanding of the potential-dependent change of the interfacial layer and structure of imidazolium-based ILs, and also the behavior of water at the electrode/electrolyte interface is necessary.

We have, therefore, investigated the interfacial molecular structure of IL ions and coadsorbed water molecules at the Pt(111) electrode surface in contact with ILs containing 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide [BMIM] as well as 1-butyl-2,3-dimethylimidazolium bis(trifluromethylsulfonyl)imide [BMMIM] cations and bis(trifluoromethylsulfonyl)imide [NTf₂] anions. For that, we have interrogated the electrode/electrolyte interfaces in situ with vibrational sum-frequency generation (SFG) spectroscopy and provided new information on the orientation of imidazolium cations at Pt surfaces in the presence of water that can be also compared to previous works^[] on imidazolium cations without a methylated C2 position (see structure details and denomination of carbon positions in Scheme ). For that reason, we mainly focus on the investigation of the potential-dependent orientation and structure of [BMMIM] cations (Scheme ) and add where needed, additional information on [BMIM] cations as a known reference system. The insights on the “electric double-layer” structure with ILs as nonconventional electrolyte that we provide are of potential interest for applications in electrocatalysis where, for example, water at the interface is needed or where IL ions can be used to stabilize reaction intermediates.

[IMAGE OMITTED. SEE PDF]

PRINCIPLES OF SFG SPECTROSCOPY

SFG spectroscopy is an inherently interface-specific method that interrogates molecular vibrations of interfacial molecules. If a molecule or a material is exposed to a weak (oscillating) electric field, the induced electric polarization P is proportional to the external electric field, but for intense light pulses, the external electric field can be very strong and the material's response is in that case not strictly linear to the electric field, but the induced electric polarization P is dependent on higher order terms that give rise to distinct nonlinear optical effects.^[] For vibrational SFG spectroscopy, two intense laser pulses are overlapped to generate a third beam with the sum frequency $${\omega _{{\rm{SF}}}} = {\omega _{{\rm{IR}}}} + {\omega _{{\rm{VIS}}}}$$ of the two incoming beams. Here, we use a frequency-fixed narrowband picosecond visible (VIS) pulse and a femtosecond IR pulse that is tunable in its frequency.^[] The SFG intensity 1 $$$\begin{eqnarray} I\left({\omega}_{\mathrm{SF}}\right) \propto {\left|{\chi}_{\mathrm{NR}}^{\left(2\right)}+{\chi}_{\mathrm{R}}^{\left(2\right)}\right|}^{2} = {\left|{\chi}_{\mathrm{NR}}^{\left(2\right)}+\sum _{q}\frac{{A}_{q}}{\omega -{\omega}_{q}+i{\mathrm{\Gamma}}_{q}}\right|}^{2},\nonumber\\[-3pt] \end{eqnarray}$$$linearly depends on the intensities of the IR and the VIS laser pulses as well as on the absolute value of second-order electric susceptibility, which can be written as the sum of a nonresonant contribution $$\chi _{{\rm{NR}}}^{( 2 )}$$ and a resonant effective second-order electric susceptibility$${\rm{\ }}\chi _{\rm{R}}^{( 2 )}$$.

In Equation (), $${A}_{q}=\ N \langle {\beta}_{q} \rangle $$ describes the amplitude of a vibrational mode $$q$$, $${\omega}_{q}$$ its resonance frequency, and its homogeneously broadened linewidth $${\mathrm{\Gamma}}_{q}$$. The amplitude $${A}_{q}$$ of a vibrational band is linearly dependent on the number of contributing molecules N as well as on their orientational average molecular hyperpolarizability $$ \langle {\beta}_{q} \rangle $$.^[] The orientational average has far reaching consequences for materials with inversion symmetry like fcc (face centered cubic) metals such as Pt or isotropic liquids and gases, where within the dipole approximation, the orientational average in the bulk material is zero and SFG is, consequently, not allowed for symmetry reasons. However, at interfaces, the bulk symmetry is necessarily broken, which gives rise to new terms of the second-order susceptibility that originate only from the interface and, thus, allow for inherently interface-specific SFG spectroscopy. As a consequence, SFG spectroscopy is essentially background free and does not require the subtraction of (strong) bulk signals.^[]

EXPERIMENTAL DETAILS

Electrochemical and spectroelectrochemical experiments were performed in a three-electrode setup. As a working electrode, we have used a Pt(111) single crystal disk with a geometric surface area of 0.78 cm² that was purchased from MaTecK (Germany). A Pt wire served as a counter electrode (99.99%, Chempur, Germany) and a RE-6 (ALS, Japan) as a reference electrode for nonaqueous media. [BMIM][NTf₂] (99%) and [BMMIM][NTf₂] (99%) ionic liquids were both purchased from IoLiTec (Germany) and were additionally purified with a molecular sieve (3 Å, 400 mg/mL) at 80°C under vacuum (∼10⁻³ mbar) for at least 16 h before each experiment. Note that we have modified the purification method that was previously suggested by Gnahm et al.^[] Before each experiment, all components were precleaned in Alconox (USA) solution, dried in an N₂ stream, and additionally cleaned in a mixture of concentrated H₂SO₄ (98%, Carl Roth, Germany) with Alnochromix (Alconox Inc., USA) for >12 h. All parts of the electrochemical cells were then rinsed with copious amounts of ultrapure H₂O and the glassware was boiled in ultrapure water three times and dried in a stream of nitrogen gas (>99.999%, Westfalen AG, Germany). All polychlorotrifluoroethylene parts of the electrochemical cell that was used for SFG spectroscopy were also thoroughly rinsed with ultrapure water, subsequent to the acid treatment, and dried in an N₂ stream but post-treated in an air plasma for 5 min instead of boiling in water. For all experiments, we have used ultrapure water from a Millipore purification system (Milli-Q Reference A+, 18.2 MΩcm, TOC < 5 ppb). The reference electrode was calibrated to the standard hydrogen electrode (SHE), where we followed the procedure that was previously published by Pavlishchuk et al. (Supporting information Figure ). ^[] The quality of the Pt(111) single-crystal electrode and its surface preparation was checked using cyclic voltammetry in a 0.1 M H₂SO₄ (96%, Suprapur from Merck) aqueous electrolyte where the well-known voltammogram of an ordered Pt(111) electrode surface was reproduced (Supporting information Figure ).^[] Cyclic voltammetry was done at 50 mV/s with a SP-150 potentiostat (Biologic, France) (Supporting information Figure ). The setup of the homemade spectroelectrochemical cell for SFG spectroscopy is described in detail elsewhere.^[] The electrolyte in the spectroelectrochemical cell was purged with CO for 30 min to saturate the ILs with CO and to eliminate O₂ as well as other dissolved gases from the ILs. The Pt(111) working electrode was flame annealed for 5 min, cooled down for additional 5 min in a reductive atmosphere that contained Ar (99.999%, Westfalen AG, Germany) and CO (99.970 %, Westfalen AG) in a ∼(3:1) ratio. Subsequently, the Pt(111) crystal was transferred into the spectroelectrochemical cell where CO was allowed to adsorb to the Pt(111) surface from the CO-saturated IL for at least 5 min while the electrode potential was kept fixed to +0.1 V versus SHE. Next, the working electrode was pressed on a 25 μm thick PTFE spacer, to form a thin-layer electrolyte between the CaF₂ window and the Pt(111) electrode. Through the use of the spacer, the thin-layer electrolyte had a well-defined thickness of 25 μm which is identical to the thickness of the PTFE spacer. We point out that the use of a 25 μm electrolyte gap drastically reduces depletion of reactants and ohmic drop effects, if the sweep rate of the electrode potential is sufficiently low.^[] In situ SFG experiments were done using a setup which was reported in detail elsewhere.^[] A VersaSTAT 3 potentiostat (Princeton Applied Research, USA) was used for electrochemical characterization. For SFG, we use a picosecond visible laser pulse with a wavelength of 804.1 nm and bandwidth of <5 cm^–1. The visible pulse is overlapped with a tunable broadband (>300 cm^–1) femtosecond IR laser pulse at the electrode/electrolyte interface in space and time with incident angles of 55° for the visible and 60° for the IR pulse versus the surface normal. The pulse energies of the visible and the IR beams were set to 22 and 10 μJ, respectively. The SFG setup was optimized using the strong SFG intensity of the CO stretching band at ∼2075 cm^–1 (Supporting information Figure ) while the electrode potential was at +0.5 V. Frequency calibration of the spectrometer was checked using the IR absorption of gaseous CO₂ in the beam path of the IR pulse (Supporting information Figure ). Once the spectrometer was optimized, the electrolyte was thoroughly purged with N₂ and the CO monolayer on the Pt(111) surface was oxidized in an anodic potential cycle from +0.5 to +1.8 V. This procedure resulted in a clean Pt(111) surface.^[] The IR center frequency was then tuned to ∼1250 cm^–1 in order to address ν(CF₃) stretching vibrations of interfacial IL anions. In addition, the visible pulse was time delayed for >300 fs with respect to the IR pulse to suppress the nonresonant contribution to the second order susceptibility according to the procedure that was first described by Lagutchev et al.^[] (Supporting information Figure ). SFG spectra with the broadband IR pulse centered at ∼1250 cm^–1 were recorded with acquisition times between 10 and 20 s, which were dependent on the SFG signal intensities, whereas the potential sweep was set to 5 and 2.5 mV/s, respectively, to reach a resolution of the electrode potential between two SFG spectra of at least 50 mV. Measurements in the frequency domain where C-H and O-H bands can be interrogated (2800–3600 cm^–1) were carried out without delaying the visible pulse and by scanning the IR frequency in at least four steps. The acquisition time of 20 s for each IR frequency and the chosen potential sweep rate of 1.25 mV/s resulted in a potential resolution of 100 mV between SFG spectra. SFG spectra that cover the C-H region only (2800 – 3100 cm^–1) were also recorded at a single IR center frequency and here again the nonresonant contribution was suppressed as explained above. This resulted in an acquisition time of 160 s and the potential sweep rate was set to 0.625 mV/s to achieve a potential resolution of 200 mV between SFG spectra.

All SFG spectra were recorded in ppp polarization (p-polarized sum-frequency, p-polarized visible, and p-polarized IR) except for the supplementary investigation at the air-liquid interface where ssp and ppp polarizations were applied. For the experiments at the air/liquid interface, a petri dish was filled with 2.2 mL of the studied ILs which also contained 0.5 M H₂O. Both spectra were normalized to a reference spectrum of a polycrystalline Au film, which was cleaned in an air plasma for 5 min prior to the experiments.

For bulk characterization, IR spectroscopy of the ILs was done at a resolution of 6 cm^–1 and a total scan time of 150 s per spectrum using a Bruker VERTEX 70 FTIR spectrometer that was equipped with a liquid N₂ cooled MCT detector. The samples were measured in an optical compartment cell where the studied ILs were pressed between two CaF₂ windows without a spacer to yield thin liquid films with a thickness of <10 μm.

EXPERIMENTAL RESULTS

Comparison of bulk and surface spectra

In order to understand the relevant vibrational bands that can be unambiguously attributed to the presence of [NTf₂] anions at the interface, we have first recorded bulk FTIR spectra as well as SFG spectra of the more easily accessible air-IL interface and compared the latter with SFG spectra that were recorded in situ at the electrode/electrolyte interface. In Figure  and , we show the FTIR spectra of [BMMIM][NTf₂] and [BMIM][NTf₂] with 0.5 M H₂O in the spectral region of 1100 to 1400 cm^–1. These vibrational spectra are dominated by strong bands of the [NTf₂] anion, although no serious difference can be observed when comparing the spectra of both ILs. SFG spectra of the air/[BMMIM][NTf₂] and air/[BMIM][NTf₂] interface in ssp and ppp polarization are presented in Figure  and , whereas the measurements with ppp polarization result in a significant lower SFG intensity compared to the ssp polarization. Therefore, the SFG intensities of spectra taken in ppp polarization at the air-liquid interface are scaled by a factor of 5 which allows for a better comparison.

[IMAGE OMITTED. SEE PDF]

We observe two dominant bands at 1137 and ∼1246 cm^–1 in both polarization combinations which have been earlier attributed to the CF₃ and SO₂ symmetric stretching vibrations of [NTf₂] anions.^[] In particular, the frequency of the CF₃ band of the interfacial species at the air-IL interface is compared to the bulk IR spectra in Figure  and , substantially blue-shifted for both cations, as indicated by the dashed line in Figure  and  at 1246 cm^–1 which corresponds to the frequency of the bulk species. The shoulder in Figure  and  at around 1208 cm^–1 is caused by the CF₃ asymmetric stretching vibration, the band which is actually dominating the bulk FTIR spectra in Figure .^[]

The corresponding asymmetric stretching vibrations have significantly lower SFG intensities in comparison to the symmetric stretching vibrations. Similar results were already reported for the air/[BMIM][NTf₂] interface by Iwahashi et al.^[] Since SFG signals are only visible for modes which are both IR and Raman active, we can conclude that the low SFG intensity is explainable by a low Raman activity of this mode.^[] On the other hand, using perfluorononanoate as an example, Tyrode et al.^[] showed with SFG spectroscopy at the vapor-liquid interface that the relative SF intensity of v_as(CH₃) increases with the tilt angle to the surface normal. Thus, it can be assumed that the [NTf₂] molecule is oriented such that the CF₃ group is nearly aligned with the surface normal axis, which is also supported by metastable atom electron spectroscopy experiments of the air/[BMIM][NTf₂] air-liquid interface by Iwahashi et al.^[] We, therefore, use the strong intensity of the symmetric CF₃ vibrational band for in situ investigation of the [NTf₂] anion at the Pt(111)/[BMIM]/[NTf₂] and Pt(111)/[BMMIM]/[NTf₂] interface. In Figure  and , we show in situ SFG spectra of Pt(111)/[BMMIM][NTf₂] and Pt(111)/[BMIM][NTf₂] interfaces where the H₂O concentration was fixed to 0.5 M. The SFG spectra are dominated by a strong band at ∼1250 cm^–1 that is attributed to ν_s(CF₃) vibrations^[] and the ν_s(SO₂) band^[] that is located at ∼1140 cm^–1. In comparison to the air/IL interface, the ν_s(SO₂) band shows an even more pronounced blue-shift in its vibrational frequency. Moreover, there are also additional SFG signals visible at frequencies below 1140 cm^–1 and above 1250 cm^–1. We relate these features in the SFG spectra to a not fully suppressed nonresonant contribution, as seen in Figure  and their appearance is weakly dependent on the alignment of the SFG setup and the exact value of the time delay that was used to suppress nonresonant contributions.

Interface spectra of imidazolium cations as a function of electrode potential

Having now established the assignment of the vibrational bands in the relevant frequency region where we can study interfacial [NTf₂] anions, we will now focus in the following part on the in situ spectroscopy of electrode/electrolyte interfaces under potential control. Complementary cyclic voltammetry experiments are presented in Supporting information Figure and revealed no major characteristic features, except the characteristic H₂O reduction feature in the cathodic region and oxide formation at anodic potentials. For a more detailed discussion, we refer to Supporting Information. In Figure  and , SFG spectra of Pt(111)/[BMMIM][NTf₂] and Pt(111)/[BMIM][NTf₂] interfaces are presented where the electrode potential was continuously changed while the acquisition of SFG spectra was synchronized to the sweep rate of the electrode potential. For a close analysis of the potential-induced changes in the SFG spectra of the electrode/electrolyte interface, we will now concentrate on the dominating ν_s(CF₃) vibrational band at ∼1250 cm^–1. Already a visual inspection, Figure  and  clearly shows that the behavior of [NTf₂] anions as reflected by the changes of ν_s(CF₃) band drastically depend on the chemical identity of the imidazolium cation ([BMMIM] vs [BMIM]).

[IMAGE OMITTED. SEE PDF]

For [BMMIM] cations, we observed a strong modulation of the SFG intensity as a function of electrode potential, while for [BMIM] cations, the SFG intensity of the ν_s(CF₃) band was only weakly dependent on the electrode potential. In order to address the changes in the SFG spectra more quantitatively, we have additionally fitted all SFG spectra in Figure  and , where we assumed model functions according to Equation () with a single Lorentzian term to accommodate the dominating ν_s(CF₃) band, while contributions of the other bands in the SFG spectra (Section 4.1) were too weak to be analyzed and were, thus, omitted during the fitting procedures.

The results of our nonlinear least square fits to the SFG spectra are shown in Figure  and , where we plot the SFG amplitude $${A}_{q}=\ N \langle {\beta}_{q} \rangle $$ as a function of the electrode potential. In case of [BMIM][NTf₂] as an electrolyte, $${A}_{q}$$ first slightly decreases during the cathodic half-cycle from the starting potential of +0.5 to -0.5 V where it starts to increase again until the reversal potential of -1.5 V is reached (Figure ).

In the anodic sweep from -1.5 to +1.5 V, the intensity remains within the experimental scatter of roughly 10% constant but decreases during the following cathodic sweep again to reach a local minimum in SFG amplitude at -0.5 V that is about 20% lower than the maximum amplitude during the anodic sweep.

The changes in SFG amplitude with electrode potential of the anion-related bands are minor for [BMIM] cations compared to the significant changes when [BMMIM] cations are present. Clearly, this already implies that the change in double-layer structure for [BMMIM][NTf₂] ILs in contact with Pt(111) is different from [BMIM][NTf₂]. Further, from an analysis of the SFG amplitudes, it is clear that the potential-induced changes are not strictly reversible but show hysteresis to some extent for [BMIM][NTf₂] electrolytes and to a large extent for [BMMIM][NTf₂] electrolytes. Interestingly, the hysteresis of the potential-induced changes of SFG amplitude from [NTf₂] anions at the Pt(111) interface in an [BMMIM][NTf₂] electrolyte is compared to an [BMIM][NTf₂] electrolyte which is not only much more pronounced, but also the increase or decrease of the amplitude for different directions of the potential sweeps is reversed (Figure  and ).

Results on water coadsorption at the interface

To investigate the behavior of H₂O at the Pt(111)/IL interface, we have recorded additional potentiodynamic SFG spectra for electrode potentials of -1.5 V and +1.9 V, where we focus spectroscopically on C-H as well as O-H stretching bands from interfacial water and [BMMIM] or [BMIM] cations at the Pt(111) electrode surface (Figures  and ). When comparing both ILs, a significant difference between the shape, the intensity, as well the position of the intensity maximum of the bands in the SFG spectra in Figures  and  can be noticed. In the case of [BMMIM][NTf₂], there is a broad band with an intensity maximum at ∼3240 cm^–1, while for [BMIM][NTf₂], a significantly blue-shifted position of the maximum of the broad band to a frequency of ∼3330 cm^–1 is noticeable. We attribute these broad features in the SFG spectra to O-H stretching vibrations ν(OH) of interfacial water molecules.

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

Here, we have to note that in the bulk solution, only narrow O-H bands with frequencies >3500 cm^–1 are observed and correspond to isolated water molecules or water molecules that form weak hydrogen bonds with anions in a double donor type structure, as reported earlier for water molecules in the bulk of an IL.^[] These structures are dominant even at very high water concentrations of about 33 M (Supporting information Figure ).

At the interface, the presence of broad O-H stretching bands that show a substantially red-shifted stretching frequency compared to free or weakly hydrogen bonded water must be attributed to the presence of an interfacial layer of water molecules within an extended hydrogen-bonded network of water molecules. This conclusion will be substantiated in more detail in the Section 5. We point out that the spectra were taken at conditions where the nonresonant contribution was not suppressed by delaying the visible versus the IR pulse, which would otherwise lead to a suppression of the broad O-H stretching bands from H-bonded interfacial water, as these have short dephasing times which is less than what is needed to suppress $$\chi _{{\rm{NR}}}^{( 2 )}$$. For that reason, there are additional strong contributions from $$\chi _{{\rm{NR}}}^{( 2 )}$$ possible that can also shape the spectra to some extent and clearly demand additional control experiments to rule out the possibility that the changes which we observe in Figures  and  are caused by changes in $$\chi _{{\rm{NR}}}^{( 2 )}$$. For that, we have performed additional experiments (Supporting information Figure ) where we have used D₂O instead of H₂O and which corroborate our conclusion that the broad bands between 3200 and 3400 cm^–1 in Figures  and  are indeed dominated by hydrogen-bonded interfacial H₂O molecules because they are absent in D₂O.

To enable a more detailed analysis of the O-H stretching bands and their potential dependence, we present the integrated SFG intensity of the broad O-H bands (I_OH) as a function of electrode potential in Figures  and  for Pt(111) in contact with [BMMIM][NTf₂] and [BMIM][NTf₂], respectively. In case of [BMMIM][NTf₂] (Figure ), the intensity of the O-H stretching band increases drastically when the electrode potential is decreased to -1.3 V during the initial cathodic sweep with a starting potential of -0.1 V (Figure ). During the subsequent anodic sweep from -1.3 to +2 V, the intensity of the O-H band decreases and plateaus at a minimum value as soon as an electrode potential of -0.2 V is established.

Overall, a slight hysteresis between anodic and cathodic potential sweeps is visible by a close analysis of Figure . This is different for the O-H intensity from interfacial water at the Pt(111)/[BMIM][NTf₂] interface, which we present in Figure . Here, the O-H intensity during the cathodic and the anodically going potential sweeps is noticeably different and shows substantial hysteresis, in particular for potentials between -0.5 and +2 V. While the increase in O-H intensity during the cathodic sweep and the decrease in the subsequent anodic sweep below -0.5 V is within the experimental scatter comparable, the O-H intensity first reaches a local minimum at -0.5 V, then increases to a broad maximum and decays to negligible intensities at electrode potentials >1.5 V. This low level in intensity remains during the sweep from +2 V to roughly -0.5 V (Figure ).

Results on [BMMIM] cations at the interface

To investigate the effects on the cation orientation, we have set the IR-Vis time delay to >300 fs, which suppresses the nonresonant component $$\chi _{{\rm{NR}}}^{( 2 )}$$ as explained in the Section 2 on SFG spectroscopy. Suppressing $$\chi _{{\rm{NR}}}^{( 2 )}$$ allows for a better analysis of the resonant components to the second-order susceptibility which we use to interrogate the C-H stretching bands from interfacial imidazolium cation without the strong interference with $$\chi _{{\rm{NR}}}^{( 2 )}$$. Figure  and  show the potentiodynamic changes of the C-H bands as a function of the electrode potential, as the structure of [BMIM] cations and the potential-induced changes have been described earlier, we are here concentrating on the spectral changes of [BMMIM] cation and substantiate our conclusion later in the Section 5 by previous works^[] on [BMIM] cation.

[IMAGE OMITTED. SEE PDF]

In Figure , three vibrational bands are clearly visible. The vibrational bands at 2868 and 2940 cm^–1 are attributable to the CH₃ symmetric stretching vibrations and the CH₃ Fermi resonance of the butyl chain,^[] while the additional band at 2970 cm^–1 has been assigned to a combination of the bands arising from the antisymmetric CH₃ stretching vibration at 2950 cm^–1 and the ν_s(N-CH₃) band at 2991 cm^–1,^[] whereas the band at 3123 cm^–1 that is weaker in SFG intensity was attributed earlier to the aromatic C-H stretching vibrations at the C4 and C5 positions of the imidazolium ring (see Scheme  for a definition of the carbon positions).^[]

The SFG amplitudes of the bands related to the [BMMIM] cations and their butyl group remain fairly constant throughout the investigated potential range, the ν_s(N-CH₃) band, while the aromatic ν(CH_ar) stretching band shows pronounced potential-induced changes. The bands in the SFG spectra were fitted also with a Lorentzian function according to Equation () and are presented in Figure . The amplitude continuously decreases between electrode potentials of -1.5 V and about 0 V, while between 0 and +1.0 V no change in amplitude can be observed. Above 1.0 V, the amplitude irreversibly decreases to 0. In particular, there is a hysteresis in SFG amplitudes noticeable for the anodic sweep.

DISCUSSION

We recall that the EDL structure of ILs at electrode interfaces is largely different from conventional electrolytes and layering of cations and anions at the interface is possible (see Section 1). The layered structure contains ions with likely some directional molecular order perpendicular to the interface. These ordered layers can also render additional molecules active for SFG spectroscopy, which were otherwise inactive due to isotropic orientational distribution, for example, within the diffuse layer of a conventional and diluted electrolyte, where the Gouy-Chapman-Stern model is still applicable. We point out that the absence of a strong electrochemical Stark tuning of the ν_s(CF₃) band can be helpful to address the EDL structure at the electrode/electrolyte interface. Within the EDL, the presence of a strong electric field directly at the electrode surface can interact with the dynamic dipole moment of interfacial species, which causes a first-order vibrational Stark tuning^[] and was shown for molecules, such as CO,^[] (bi)sulfate,^[] or nitric oxide,^[] that were directly adsorbed to the electrode surface. On the other hand, Tong et al.^[] investigated the electrooxidation of formic acid on Pt(100). Here, they reported weakly adsorbed formic acid, which is accompanied by a blue-shift in frequency in the positive scan until a threshold potential was reached. In the negative scanning, no Stark shift was observed, where formic acid molecules are expected to directly approach the surface. In both cases, there was no direct contact between formic acid and metal assumed.^[] Because the frequency changes of the ν_s(CF₃) band (Figure ) as a function of electrode potential are negligible and a pronounced Stark tuning is clearly absent, we propose that the band originates from interfacial [NTf₂] anions which are not directly adsorbed to the Pt surface, but are located at the vicinity of the interface. This hypothesis is further substantiated below where we demonstrate that an interfacial layer of water molecules exists that can hinder direct adsorption of IL ions. That is because they would otherwise lack SFG activity (see Section 2). Consistent results on the absence of a strong Stark tuning were already reported for [DCA] anions at Pt(poly)/[EMIM][DCA]^[] and Pt(poly)/[BMIM][DCA]^[] interfaces, where no Stark tuning was detectable for the ν_as(CN) stretching vibration. At negative potentials, the Pt surface gets negatively charged, which forces the [NTf₂] anions to leave the interface due to electrostatic repulsion. This is different when the Pt surface gets positively charged, where the anions get attracted to interface and the SFG intensity of the ν_s(CF₃) band in Figure  rises again.

Assuming that the [NTf₂] anions are not directly adsorbed to the platinum surface, there are two possibilities on how interfacial molecules can interact with the Pt(111) surface directly: (i) The imidazolium cations could be adsorbed at the interface in a first layer throughout the entire potential window from -1.5 to +1.5 V. (ii) H₂O has to be considered as an additional component of the electrical double layer that can have an effect on the orientation and coverage of ions at the interface as it competes for adsorption sites but can also screen the electric field at the interface.

We first want to discuss the role of the cation in the EDL. Due to the fact that no Stark tuning is noticeable for C-H stretching bands over the whole potential range, a direct adsorption of [BMMIM] on the Pt(111) surface is excluded. This conclusion is consistent with similar results that were already observed for [BMIM] cations at the Pt(111) interface.^[] For a more detailed analysis of the cation orientation, especially of the ring, the aromatic C-H vibrational bands of the C4-H and C5-H bonds of the imidazolium ring can be used as nicely shown in previous works on other imidazolium cations and which was first introduced by Baldelli and co-workers.^[] For the analysis, it is assumed that the imidazolium ring can be approximated by a C_2v symmetry and that its orientation to the surface normal of the Pt(111) surface can be simplified by the tilt angle θ in Figure  as well as by a rotation along an axis that lies within the ring plane and points through the C2 position (Figure ).^[]

[IMAGE OMITTED. SEE PDF]

Qualitatively, a tilt angle θ of 90° is equivalent to a ring orientation parallel to the surface, whereas a θ of 0° corresponds to an upright orientation, with the SFG intensity of the vibrational bands being smaller at a higher tilt angle: That is mainly because of the dipole selection rules on metal surfaces where dynamic dipole moments of vibrational modes parallel to the surface plane of the metal are zero because of the induced image dipole on the metal site of the interface. As a consequence, molecular vibrations with a dynamic dipole moment parallel to the metal surface, for example, for the aromatic stretching vibrations from an imidazolium ring that is lying parallel to the surface, are IR as well as SFG inactive. In this work, we are following the line of arguments from previous works on different imidazolium cations and propose that the changes in the SFG intensity of the aromatic C-H band at 3124 cm^–1 in Figure  can be associated to a reorientation of the imidazolium ring that is caused by the applied electrode potential. Consequently, for [BMMIM], a more perpendicular orientation to the surface is established at very negative potentials because here the SFG intensity is high while the SFG signal of the aromatic C-H band is close to zero at high anodic potentials, which we can associate to an increased tilt angle with the imidazolium ring more parallel to the Pt surface. We note that the drastic changes in aromatic C-H bands SFG intensity are accompanied by almost negligible changes in the SFG intensity of other C-H bands which appear unaffected by electrode potential (Figure ). We will now compare our results with previous studies on the reorientation of IL cations at electrode/electrolyte interfaces, which reported on a different potential dependence. For instance, Baldelli^[] reports on in situ SFG experiments at the Pt/[BMIM][BF₄] interface that were done under dry vacuum conditions and, thus, with ILs that had a negligible water content. In this previous work, an orientation where the [BMIM] cation lies more parallel to the surface at cathodic potentials and gets repelled with increasing electrode potential to establish a more perpendicular orientation was proposed. This is similar to the reorientation of [BMIM] cations at Au electrodes as shown by Motobayashi et al.^[] who have used surface-enhanced infrared absorption spectroscopy (SEIRAS) and have done their experiments also under dry vacuum conditions. For that, Motobayashi et al.^[] monitored the intensity increase of C-H stretching vibration at the imidazolium ring during an anodic sweep, and used the dipole selection rules explained above to conclude that the imidazolium cations change from a flat-lying orientation with negligible band intensity to a more upright (or vertical) orientation when the electrode potential was increased.^[] Besides electrode-ion interactions, where no significant effect of the electrode material, for example, Au versus Pt on the orientational changes of the [BMIM] cation was observed, ion-ion interactions have to be considered as well. In the same study, Motobayashi et al.^[] have additionally investigated the effects of the choice of the anion for which they compared data for [BF₄] and [NTf₂] cations. The data did not show a substantial difference in the trend of the cation reorientation with electrode potential but instead demonstrated that the hysteresis in reorientation was absent in the presence of [BF₄], while it was substantial for [NTf₂]. This was attributed to the higher steric hindrance of the bulky [NTf₂] anion compared to the smaller [BF₄], where the authors proposed a higher overpotential and, thus, hysteresis for the restructuring process in the case of [NTf₂].^[] In contrast to the studies described above, which were all done under dry vacuum conditions and, thus, in the absence of water, recent in situ SFG experiments^[] at the Pt/[EMIM][BF₄] interface in the presence of 0.5 M H₂O reported on a more upright orientation of the imidazolium ring when the electrode potential was negative or positive with respect to the potential of zero charge (pzc), while a more parallel orientation with weak SFG intensities was realized at the pzc. Clearly, the presence of H₂O has a significant effect on the potential dependence of the cation reorientation and we will elaborate more on the exact role of H₂O at the interface below. First, we will discuss the role of the cation's electronic structure which additionally needs to be taken into account when addressing the cation reorientation at the interface. Indeed, in a previous study by Motobayashi et al.^[] it was demonstrated for the Au/IL interface that [BMPyrr] cations behaved differently compared to [BMIM]. In fact, SEIRAS at the Au/[BMPyrr][NTf₂] interface showed that the ν(CH₂)_pyrr band intensity decreased during the anodic sweep. Here, the authors also deploy the argument of the dipole selection rules on metal surfaces, from which they rule out the possibility of a reorientation from flat to a more vertical orientation with increasing electrode potential. This is clearly different from the case of [BMIM] and the difference was attributed to the higher electronic polarization of imidazolium cations within the interfacial electric field as compared to pyrrolidinium cations where the charges at the ring are expected to be more localized.^[] These observations and their justification can also shed some light on the differences between [BMIM] and [BMMIM] cations, which we reported in our study. Although the investigated [BMMIM] belongs to the imidazolium cations, we see a similar potential dependence as previously reported for [BMPyrr] at the Au/[BMPyrr][NTf₂] interface. This can now be explained as follows: Hunt et al.^[] calculated the partial charges at the C2 position of [BMMIM] and [BMIM] cations which were almost twofold higher for [BMMIM], as compared to [BMIM] and revealed that the removed positive partial charge of the H atom is not compensated by the electron-donating (+I) effect of the methyl group at the C2 position of the imidazolium ring. This leads to a localized charge distribution at the imidazolium ring of [BMMIM] which can, thus, cause an orientation change from a more upright orientation to a flat-lying orientation similar to what was reported for [BMPyrr][NTf₂] (see above). Moreover, MD simulations by Wang et al.^[] demonstrated that [EMMIM] cations exhibit a more horizontal orientation to the Au surface than [EMIM] cations, which was rationalized by increased electrode-ion interactions of [EMMIM][NTf₂] in contrast to [EMIM][NTf₂]. These conclusions from simulation have been experimentally confirmed by in situ STM and AFM experiments that were done by Bingwei Mao and co-workers.^[] In comparison to [BMIM] cations, it is therefore reasonable to assume that [BMMIM] cations can also form structures at electrode potential that are largely different from other imidazolium cations like [BMIM] and which we attribute to the more localized charges at the imidazolium ring of [BMMIM] and, consequently, to increased electrode-ion interactions which are also a function of the H₂O concentration. Clearly, even a rather minor chemical modification of imidazolium cations like adding a methyl group at the C2 position (Scheme ) can have significant influence on the potentiodynamic orientation of the cation at the interface as well as on the performance of the IL in applications such as in electrocatalysis.^[]

As we have discussed above, cations and anions of both ILs [BMMIM][NTf₂] and [BMIM][NTf₂] are not directly adsorbed to the Pt(111) electrode surface in the potential range covered by this study. Further, we have shown that water molecules are present at the interface and form a layer of hydrogen-bonded molecules at the interface whereas isolated water molecules without strong hydrogen bonds dominate the bulk. From these observations, we can draw the conclusion that hydrogen-bonded water must be adsorbed to the Pt(111) surface in form of a first layer adjacent to the interfacial IL ions. This conclusion is supported by theoretical investigations by Feng et al.^[] as well as by Kobayashi et al.,^[] who have demonstrated that H₂O can accumulate within a subnanometer distance from charged and even uncharged interfaces when imidazolium-based ILs are present. The infusion of water into the interfacial layers can be driven by an entropy gain.^[] It was also shown that three factors exist which can influence the water distribution at the interface in humid ILs. (i) The interactions between water dipoles and the electric field, (ii) the interaction of H₂O with the surrounding IL ions, and (iii) the availability of free space or voids near the electrode surface that can be filled by water to increase the interfacial entropy.^[] In contrast to that, water accumulation was only seen at negatively charged graphite/Li[NTf₂] interfaces, whereas water gets repelled from the first interfacial layer at positive potentials and a significant increase of anion adsorption is observed.^[] This is in contrast to in situ SEIRAS results from the Au/[BMIM][NTf₂] interface, where water accumulation only occurred at positive charged interfaces due to strong interaction of water with the [NTf₂] anions.^[] Using a Pt electrode, we revealed H₂O accumulation at the Pt/IL interface when [BMIM] was present at both positive as well as negative electrode potentials (Figure ), while water was found at the interface only at negative potentials in case of [BMMIM] (Figure ). The different results compared to previous literature can be, therefore, attributed to the different interplay of the above-mentioned factors which also seem to be significantly influenced by the cation orientation that is dependent on the degree of charge delocalization and, thus, the cations ability to screen the interfacial electric field.

At very negative potentials, we show a more upright orientation of the [BMMIM] cation, this configuration can introduce additional gaps or voids between the cation-rich layer and the Pt(111) surface that are likely filled by H₂O molecules while the system gains entropy. With increasing electrode potential, the ring is oriented increasingly more parallel to the surface and forms a close-packed layer, which can cause a decrease in the density of voids for the incorporation of water. This is consistent with the decrease in the O-H intensity from interfacial H₂O that we have observed (Figure ). Above +1.0 V, interfacial H₂O is likely consumed on the Pt surface as hydroxyl and other oxygen species form a new first interfacial layer, as suggested by Scheme . Here, we also note that this process also happens for CO oxidation reactions that were used to prepare clean Pt(111) surfaces in ILs. However, compared to aqueous electrolytes, large overpotentials for surface and CO oxidation do exist.

[IMAGE OMITTED. SEE PDF]

Compared to [BMMIM], [BMIM] also shows H₂O accumulation in the anodic region up to +1.0 V, which is expected to result in a less tilted cation structure in the negative and positive potential regions, as observed for the Pt(111)/[EMIM][BF₄] interfaces with 0.5 M H₂O by Kemna et al.^[] Recently, Zhang et al.^[] investigated the water structure at the electrified Pd/H₂O interface with ATR-SEIRAS, where they reveal surface H₂O as composed of “ice-like” (∼3200 cm^–1), “liquid-like” (∼3400 cm^–1), and small clusters of water molecules (∼3600 cm^–1). We can therefore say that intermolecular interactions of different cations lead to a change of interfacial H₂O structure, whereas [BMMIM][NTf₂] stabilizes water molecules in a more ordered H-bonded network, and water molecules at Pt(111) surfaces in contact with [BMIM][NTf₂] are more disordered with a higher center frequency of the broad O-H stretching band that shifted from ∼3200 to >3300 cm^–1. Similar results of a liquid-like interfacial H₂O structure is also reported for the Pt(111)/[EMIM][BF₄] interface.^[]

As we now have a detailed description of the interfacial structure at different potentials, we turn to the observed hysteresis in the changes of the interfacial [NTf₂] anions (Figure ). Several authors proposed that the lateral diffusion of ions in the layered structure is significantly lower than the diffusion in bulk IL, their response to a change in electrode potential is, thus, slow and can contribute to the observed hysteresis, which becomes indeed more significant with higher sweep rates.^[] In previous works, it was argued that this effect originates from nonequilibrium state during restructuring of the interface, which was observed by analyzing the capacitance at Au/[BMIM][OTf] interfaces.^[] Another origin for hysteresis in the potential dependence is the existence of an energy barrier for interface restructuring as was reported by Zhou et al.^[] who studied the anion ad- and desorption processes at Pt(poly)/[BMIM][OTf] interfaces with as received [BMIM][OTf] using SFG spectroscopy. For that, the authors have recorded SFG spectra at three different potentials after holding each potential for 20 min to establish a new equilibrium state. Similar hysteresis of the potential-dependent restructuring of vacuum-dried [BMIM][NTf₂], but with a polycrystalline Au electrode was shown previously by SEIRAS.^[] Changes in hysteresis were visible with a magnitude larger than in our experiments (2, 25, 500 mV/s).^[] Since we measured below 5 mV/s in all experiments, it can be assumed that the influence of the former hysteresis has no significant effect on the structure of our spectra and we can make a qualitative analysis. Voroshylova et al.^[] used MD simulation to investigate the Au(111)/[BMIM][PF₆] interface in terms of potential scanning direction on the interfacial structure. Here, hysteresis in the capacitance-potential curves was found, indicating the coexistence of two types of structures that differ in how counterions (over)compensate the surface charge, whereas high energy barriers for a phase transition from one ordered structure to another slow down the interfacial restructuration and lead to the observed hysteresis.^[] Real-time measurements show that the response to the applied potential is >10 s,^[] and with the help of laser-induced temperature jump experiments Sebastián et al.^[] observed a clear hysteresis for Pt(111) electrodes in [EMMIM][NTf₂] with 50 ppm H₂O, indicating the need of significant overpotential to change the layered structure. They also compared the hysteresis at other Pt(hkl)/IL interfaces with Au(hkl)/IL interfaces, where the Au electrodes induced a much smaller hysteresis. Consequently, this suggests a higher energy barrier to change the layered structure of ILs at Pt as compared to Au electrode surfaces.^[] The reason for increased overpotential for changing the layered structure in the case of [EMMIM] can be rationalized as follows. A CH₃ group at the C2 position (Scheme ) of the imidazolium ring increases the ability of more localized Coulomb interactions in the case of [BMMIM][NTf₂] but weakens the ability to form H-bonds as already shown by Hunt et al.^[] due to the increased positive charge at the C2 position. Moreover, the authors argue that methylation reduces the ability to rotate the alkyl chain and, thus, leads to a reduced configurational entropy.^[] The electron redistribution due to the methyl group at the C2 position also causes an increased anion-cation interactions giving rise to an higher molecular order as well as tighter packing of [BMMIM][NTf₂] ions in comparison to [BMIM][NTf₂].^[] AFM experiments by Liu et al.^[] indeed revealed a more layered structure for [EMMIM][NTf₂] in contrast to [EMIM][NTf₂] ILs at Au surfaces which the authors have ascribed to higher cation-anion interactions in [EMMIM][NTf₂]. This more layered structure and increased cation-anion interactions increase the overpotentials for EDL restructuration for [EMMIM] compared to [EMIM].

Considering the above discussed results from previous works, we can now make the following statements regarding the hysteresis that we observe in our study. In the case of [BMMIM][NTf₂], the initial decrease of the SFG intensity from [NTf₂] anions in Figure  is probably due to nonequilibrium condition at the interface, since no sufficient equilibration time was given during the potential sweep. Intensity changes below -0.5 V and above +0.5 V indicate hysteresis caused by an energy barrier that necessarily has to be crossed in order for the [NTf₂] anions to reorient at the interface, whereas for interfacial water and the imidazolium cations, the very small hysteresis indicates no large energy barrier for restructuring. Consequently, we propose that no significant exchange of cation-rich layers takes place, but only anions or water exhibit some mobility at the interface and screen the surface charges by moving away from or closer to the Pt(111) surface which depends on the applied potential. The strong hysteresis in the cathodic region is likely due to the high activity of Pt surfaces for hydrogen evolution reactions (HER), which is accompanied by consumption of H₂O and the reduction currents at electrode potentials of ← 1.25 V, seen in cyclic voltammetry ().

In contrast to the [BMMIM][NTf₂] interface with Pt(111), there is a reduced hysteresis of the potential dependence of [NTf₂] anion in Figure  when Pt(111) was immersed in [BMIM][NTf₂] electrolytes. Here, there are only very small intensity differences in SFG spectra visible, which indicate little mobility or restructuring of the anions when the electrode potential was changed. In addition, the presumably less ordered interface makes it difficult for the water to enter or leave the interface dynamically which we propose may be one possible origin for the observed strong hysteresis in Figure .

The results on the potentiodynamic restructuring of the EDL at Pt(111)/IL interfaces, can additionally be used to describe the influence on electrochemical reactions, like for example CO₂ reduction reactions. Previous publications on CO₂ reduction reactions at the electrode/IL interfaces revealed that H₂O plays a crucial role in lowering CO formation potentials.^[] At a Pt(poly)/[BMIM][NTf₂] interface and a H₂O concentration of 0.5 M, a CO formation potential of -0.4 V versus SHE was observed.^[] A comparison with the in situ SFG spectra in the OH region of Pt(111)/[BMIM][NTf₂] (Figure ) shows a comparatively low H₂O concentration at the interface, leading to the assumption that only a small interfacial water layer is required for the reduction of the CO formation potential.

CONCLUSIONS

We have applied in situ SFG spectroscopy to investigate the role of the chemical nature of imidazolium cations on the potential-dependent reordering of the EDL at Pt(111) electrodes in [BMIM][NTf₂] and [BMMIM][NTf₂]. For that, we report on ν(CF₃), ν(OH), and ν(CH_ar) bands to track the changes of interfacial anions, water, and cations as a function of electrode potential. From our results, we can draw the following conclusions, as summarized in the schematic representation in Scheme : Since no significant Stark tuning was observed for both the cations and the anions, while the SFG intensity was nonzero, some preferential orientation of the ions at the interface exists with the ions in close vicinity of the Pt(111) surface without being specifically adsorbed. Interestingly, we reveal the presence of interfacial H₂O for [BMMIM][NTf₂] and [BMIM][NTf₂] with the water molecules in a hydrogen-bonded network that gives rise to broad and highly red-shifted O-H stretching bands as opposed to the narrow and less red-shifted O-H bands of water in the bulk IL. Strong cation-anion interactions of [BMMIM][NTf₂] in comparison to [BMIM][NTf₂] lead to the formation of a well-ordered and layered structure with a significantly more horizontal orientation of the imidazolium ring for anodic electrode potentials and more perpendicular orientation for cathodic potentials. The different reorientations are primarily a consequence of the different ability of [BMMIM] and [BMIM] cations to screen the int0erfacial electric field. This is mainly caused by the differently delocalized positive charges, which significantly more delocalized for [BMIM] compared to [BMMIM]. In addition, accumulation of H₂O at the interface was shown in the cathodic region for [BMMIM][NTf₂] and in both the anodic and cathodic potential regions for [BMIM][NTf₂], which is possibly directly related to the different cation orientation at the interface and the resulting space demand that can dominate the incorporation of water into the interface to a large extent. We propose that a higher energy barrier for EDL restructuring is present for [BMMIM][NTf₂] and gives rise to a pronounced hysteresis of the reorientation of interfacial [NTf₂] anions. Small changes in the cation structure, thus, lead to a significant change in the potential-dependent EDL structure and may have a significant influence in electrocatalysis.

[IMAGE OMITTED. SEE PDF]

ACKNOWLEDGMENTS

We gratefully acknowledge the funding of project BR4760/3-1 and BR4760/3-2 from the Deutsche Forschungsgemeinschaft (DFG).

CONFLICT OF INTEREST

The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

D. S. Silvester, R. Jamil, S. Doblinger, Y. Zhang, R. Atkin, H. Li, J. Phys. Chem. C 2021, 125, [eLocator: 13707].

H. Sakaebe, H. Matsumoto, K. Tatsumi, Electrochim. Acta 2007, 53, 1048.

X. Wang, M. Salari, D.‐E. Jiang, J. Chapman Varela, B. Anasori, D. J. Wesolowski, S. Dai, M. W. Grinstaff, Y. Gogotsi, Nat. Rev. Mat. 2020, 5, 787.

K. Ma, C. Zhang, C. E. Woodward, X. Wang, Electrochim. Acta 2018, 289, 29.

A. Lahiri, G. Pulletikurthi, F. Endres, Front. Chem. 2019, 7, 85.

S. Liu, L. Zhou, P. Wang, F. Zhang, S. Yu, Z. Shao, B. Yi, ACS Appl. Mater. Interfaces 2014, 6, 3195.

B. Ratschmeier, B. Braunschweig, J. Phys. Chem. C 2021, 125, [eLocator: 16498].

Z. Jusys, J. Schnaidt, R. J. Behm, J. Chem. Phys. 2018, 150, [eLocator: 41724].

Y. Zhang, M. W. Rutland, J. Luo, R. Atkin, H. Li, J. Phys. Chem. C 2021, 125, 3940.

P. Bonhôte, A.‐P. Dias, N. Papageorgiou, K. Kalyanasundaram, M. Grätzel, Inorg. Chem. 1996, 35, 1168.

F. Endres, O. Höfft, N. Borisenko, L. H. Gasparotto, A. Prowald, R. Al‐Salman, T. Carstens, R. Atkin, A. Bund, S. Zein El Abedin, Phys. Chem. Chem. Phys. 2010, 12, 1724.

A. Kemna, N. García Rey, B. Braunschweig, ACS Catal. 2019, 9, 6284.

M. Gouy, J. Phys. Theor. Appl. 1910, 9, 457.

D. L. Chapman, Phil. Mag. 1913, 25, 475.

O. Stern, Z. Elektrochem. 1924, 508.

K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. Int. Ed. 2008, 47, 8731.

K. Fumino, A. Wulf, R. Ludwig, Angew. Chem. Int. Ed. 2008, 47, 3830.

R. P. Matthews, T. Welton, P. A. Hunt, Phys. Chem. Chem. Phys. 2015, 17, [eLocator: 14437].

F. Tang, T. Ohto, T. Hasegawa, M. Bonn, Y. Nagata, Phys. Chem. Chem. Phys. 2017, 19, 2850.

X. Gong, A. Kozbial, F. Rose, L. Li, ACS Appl. Mater. Interfaces 2015, 7, 7078.

Y.‐L. Wang, A. Laaksonen, M. D. Fayer, J. Phys. Chem. B 2017, 121, 7173.

C. Aliaga, C. S. Santos, S. Baldelli, Phys. Chem. Chem. Phys. 2007, 9, 3683.

M. V. Fedorov, N. Georgi, A. A. Kornyshev, Electrochem. Commun. 2010, 12, 296.

M. V. Fedorov, A. A. Kornyshev, J. Phys. Chem. B 2008, 112, [eLocator: 11868].

M. V. Fedorov, A. A. Kornyshev, Electrochim. Acta 2008, 53, 6835.

N. Georgi, A. A. Kornyshev, M. V. Fedorov, J. Electroanal. Chem. 2010, 649, 261.

K. Kirchner, T. Kirchner, V. Ivaništšev, M. V. Fedorov, Electrochim. Acta 2013, 110, 762.

J. P. de Souza, Z. A. H. Goodwin, M. McEldrew, A. A. Kornyshev, M. Z. Bazant, Phys. Rev. Lett. 2020, 125, [eLocator: 116001].

G.‐B. Pan, W. Freyland, Chem. Phys. Lett. 2006, 427, 96.

S. Baldelli, Acc. Chem. Res. 2008, 41, 421.

S. Rivera‐Rubero, S. Baldelli, J. Phys. Chem. B 2004, 108, [eLocator: 15133].

S. Baldelli, J. Phys. Chem. B 2005, 109, [eLocator: 13049].

A. Kemna, B. Braunschweig, J. Phys. Chem. Lett. 2020, 11, 7116.

W. Zhou, Y. Xu, Y. Ouchi, ECS Trans. 2013, 50, 339.

W. Zhou, S. Inoue, T. Iwahashi, K. Kanai, K. Seki, T. Miyamae, D. Kim, Y. Katayama, Y. Ouchi, Electrochem. Commun. 2010, 12, 672.

B. Ratschmeier, A. Kemna, B. Braunschweig, ChemElectroChem 2020, 7, 1765.

B. A. Rosen, A. Salehi‐Khojin, M. R. Thorson, W. Zhu, D. T. Whipple, P. J. A. Kenis, R. I. Masel, Science (New York, N.Y.) 2011, 334, 643.

R. W. Boyd, Nonlinear optics, Academic Press, Amsterdam, 2008.

Y. R. Shen, Principles of nonlinear optics, Wiley, Hoboken, NJ, 2003.

G. Q. Lu, A. Lagutchev, D. D. Dlott, A. Wieckowski, Surf. Sci. 2005, 585, 3.

A. Lagutchev, S. A. Hambir, D. D. Dlott, J. Phys. Chem. C 2007, 111, [eLocator: 13645].

L. J. Richter, T. P. Petralli‐Mallow, J. C. Stephenson, Opt. Lett. 1998, 23, 1594.

O. Isaienko, E. Borguet, Opt. Express 2012, 20, 547.

G. L. Richmond, Chem. Rev. 2002, 102, 2693.

C. D. Bain, Faraday Trans. 1995, 91, 1281.

A. G. Lambert, P. B. Davies, D. J. Neivandt, Appl. Spectrosc. Rev. 2005, 40, 103.

Y. R. Shen, Nature 1989, 337, 519.

M. Gnahm, D. M. Kolb, J. Electroanal. Chem. 2011, 651, 250.

M. A. Montiel, J. Solla‐Gullón, V. Montiel, C. M. Sánchez‐Sánchez, Phys. Chem. Chem. Phys. 2018, 20, [eLocator: 19160].

V. V. Pavlishchuk, A. W. Addison, Inorg. Chim. Acta 2000, 298, 97.

B. Braunschweig, P. Mukherjee, D. D. Dlott, A. Wieckowski, J. Am. Chem. Soc. 2010, 132, [eLocator: 14036].

N. García Rey, E. Weißenborn, F. Schulze‐Zachau, G. Gochev, B. Braunschweig, J. Phys. Chem. C 2019, 123, 1279.

P. Sebastian, M. Tułodziecki, M. P. Bernicola, V. Climent, E. Gómez, Y. Shao‐Horn, J. M. Feliu, ACS Appl. Energy Mater. 2018, 1, 4617.

K. Noack, P. S. Schulz, N. Paape, J. Kiefer, P. Wasserscheid, A. Leipertz, Phys. Chem. Chem. Phys. 2010, 12, [eLocator: 14153].

J. C. Lassègues, J. Grondin, R. Holomb, P. Johansson, J. Raman Spectrosc. 2007, 38, 551.

T. Iwahashi, Y. Sakai, D. Kim, T. Ishiyama, A. Morita, Y. Ouchi, Faraday discuss. 2012, 154, 289.

I. Rey, J. C. Lassègues, J. Grondin, L. Servant, Electrochim. Acta 1998, 43, 1505.

V. H. Paschoal, L. F. O. Faria, M. C. C. Ribeiro, Chem. Rev. 2017, 117, 7053.

E. Tyrode, C. M. Johnson, M. W. Rutland, J. P. R. Day, C. D. Bain, J. Phys. Chem. C 2007, 111, 316.

T. Iwahashi, T. Nishi, H. Yamane, T. Miyamae, K. Kanai, K. Seki, D. Kim, Y. Ouchi, J. Phys. Chem. C 2009, 113, [eLocator: 19237].

L. Cammarata, S. G. Kazarian, P. A. Salter, T. Welton, Phys. Chem. Chem. Phys. 2001, 3, 5192.

Y. Danten, M. I. Cabaço, M. Besnard, J. Mol. Liq. 2010, 153, 57.

T. Köddermann, C. Wertz, A. Heintz, R. Ludwig, Angew. Chem. Int. Ed. 2006, 45, 3697.

C. Aliaga, S. Baldelli, J. Phys. Chem. B 2006, 110, [eLocator: 18481].

R. A. MacPhail, H. L. Strauss, R. G. Snyder, C. A. Elliger, J. Phys. Chem. 1984, 88, 334.

R. G. Snyder, H. L. Strauss, C. A. Elliger, J. Phys. Chem. 1982, 86, 5145.

B. D. Fitchett, PhD Thesis, The University of Utah (USA), 2006.

G. R. Bell, Z. X. Li, C. D. Bain, P. Fischer, D. C. Duffy, J. Phys. Chem. B 1998, 102, 9461.

K. M. Dieter, C. J. Dymek, N. E. Heimer, J. W. Rovang, J. S. Wilkes, J. Am. Chem. Soc. 1988, 110, 2722.

D. K. Lambert, Electrochim. Acta 1996, 41, 623.

M. J. Weaver, Appl. Surf. Sci. 1993, 67, 147.

B. Braunschweig, P. Mukherjee, J. L. Haan, D. D. Dlott, J. Electroanal. Chem. 2017, 800, 144.

G. Zwaschka, M. Wolf, R. K. Campen, Y. Tong, Surf. Sci. 2018, 678, 78.

A. Rodes, V. Climent, J. M. Orts, J. M. Pérez, A. Aldaz, Electrochim. Acta 1998, 44, 1077.

Y. Tong, K. Cai, M. Wolf, R. K. Campen, Catalysis Today 2016, 260, 66.

K. Motobayashi, K. Minami, N. Nishi, T. Sakka, M. Osawa, J. Phys. Chem. Lett. 2013, 4, 3110.

K. Motobayashi, Y. Shibamura, K. Ikeda, Electrochem. Commun. 2019, 100, 117.

P. A. Hunt, J. Phys. Chem. B 2007, 111, 4844.

Y. Wang, Y. Sun, Y. Dong, G. Tian, J. Phys. Chem. B 2021, 125, 3677.

S. Liu, J. Peng, L. Chen, P. Sebastián, J. M. Feliu, J. Yan, B. Mao, Electrochim. Acta 2019, 309, 11.

G. Feng, X. Jiang, R. Qiao, A. A. Kornyshev, ACS nano 2014, 8, [eLocator: 11685].

T. Kobayashi, A. Kemna, M. Fyta, B. Braunschweig, J. Smiatek, J. Phys. Chem. C 2019, 123, [eLocator: 13795].

J. Vatamanu, O. Borodin, J. Phys. Chem. Lett. 2017, 8, 4362.

K. Motobayashi, M. Osawa, Electrochem. Commun. 2016, 65, 14.

S. Gilman, J. Chem. Phys. 1964, 68, 70.

Y.‐J. Zhang, Z.‐F. Su, J.‐F. Li, J. Lipkowski, J. Phys. Chem. C 2020, 124, [eLocator: 13240].

S. Makino, Y. Kitazumi, N. Nishi, T. Kakiuchi, Electrochem. Commun. 2011, 13, 1365.

N. Nishi, Y. Hirano, T. Motokawa, T. Kakiuchi, Phys. Chem. Chem. Phys. 2013, 15, [eLocator: 11615].

M. Sha, G. Wu, Q. Dou, Z. Tang, H. Fang, Langmuir 2010, 26, [eLocator: 12667].

Q. Dou, M. Sha, H. Fu, G. Wu, ChemPhysChem 2010, 11, 2438.

T. R. Gore, T. Bond, W. Zhang, R. W.J. Scott, I. J. Burgess, Electrochem. Commun. 2010, 12, 1340.

I. V. Voroshylova, H. Ers, B. Docampo‐Álvarez, P. Pikma, V. B. Ivaništšev, M. N. D. S. Cordeiro, J. Phys. Chem. Lett. 2020, 11, [eLocator: 10408].

A. Uysal, H. Zhou, G. Feng, S. S. Lee, S. Li, P. Fenter, P. T. Cummings, P. F. Fulvio, S. Dai, J. K. McDonough, Y. Gogotsi, P. Fenter, J. Phys. Chem. C 2014, 118, 569.

P. Sebastián, A. P. Sandoval, V. Climent, J. M. Feliu, Electrochem. Commun. 2015, 55, 39.

P. Sebastián, V. Climent, J. M. Feliu, Electrochem. Commun. 2016, 62, 44.

P. Sebastián, E. Gómez, V. Climent, J. M. Feliu, Electrochim. Acta 2019, 311, 30.

A. Hailu, S. K. Shaw, Energy Fuels 2018, 32, [eLocator: 12695].